Providing Optimal Model for Portfolio Optimization by Using Simulated Annealing Algorithm (SA) and Hybrid Algorithm of the Simulated Annealing Algorithm (SA) and Tabu Search Algorithm (TS)
Providing Optimal Model for Portfolio Optimization by Using Simulated Annealing Algorithm (SA) and Hybrid Algorithm of the Simulated Annealing Algorithm (SA) and Tabu Search Algorithm (TS)
Sara Ziyari1
1) M. A. in Accounting -
IOI Code
:
XYZH-ZADFZ
Publication :
2nd International Conference on Management and Humanities(2icmhconf.com)
Abstract :
As a result of lots of criticism to the mean–variance model of Markowitz due to it’s unrealistic assumptions, by adding cardinality constrained, in this article we replace a quadratic mixed-integer problem with the classical quadratic optimization model in order to capture market realism. Simulated annealing algorithm and hybrid algorithm of simulated annealing (SA) and tabu search (TS) are applied to solve the model used in this study. The main purpose of this article is to evaluate the efficiency of these algorithms in providing high quality solutions in a reasonable time for solving portfolio optimization problem. The results of evaluating the efficient frontier of mean-variance model and the efficient frontier obtained by using SA and TS-SA algorithm, for 50 companies listed in Tehran Stock Exchange for 30 months ending 22th September 2014, indicates that both algorithms have relatively good ability to solve cardinality constrained portfolio optimization problem. Additionally, comparing the percentage error of the results of the algorithms from the standard efficient frontier shows slight error (high quality) and slight computational time of TS-SA algorithm and consequently it’s greater efficiency compared with SA algorithm.
Keywords : simulated annealing algorithm hybrid algorithm of simulated annealing and tabu search portfolio optimization